CWRF 2016 Workshop ESRF Grenoble France June 21

CWRF 2016 Workshop ESRF, Grenoble, France June 21 - 24, 2016 June 2, 2016 : 10 th anniversary of the 1 st beam production by SOLEIL 10 YEARS OF OPERATION WITH THE SOLEIL RF SYSTEMS Experience, upgrades, R&D’s P. MARCHAND 1

SOLEIL Booster RF system Ø En : 100 Me. V 2. 75 Ge. V (rep. 3 Hz) ; VRF : 0. 2 1 MV @ 352 MHz Ø 1 x 5 -cell Cu cavity (CERN LEP) Ptot : 25 k. W (Pdis : 20 k. W, Pbeam : 5 k. W) Ø 1 x solid state amplifier (SSA) 35 k. W CW @ 352 MHz (developed in house) Cavity in the BO ring BO RF room (amplifier with LLRF & control) ~ 58 000 running hours over 10 years and only 3 short downtimes in operation (~ 0. 5 10 -4 overall). The 35 k. W SSA (~ 150 RF modules of 300 W) still exhibits 100% operational availability. ~ 1 module failure / year, without impact on the operation, thanks to the modularity and redundancy. 2

Booster RF upgrade plan The objective is to improve the injection efficiency in the low-a operation mode by a operation factor of ~ 2 for radiation safety reasons. That requires increasing VRF from 1 MV up to 3 MV in order to achieve the proper bunch length. MV Ø Install our available spare cavity in the BO ring and power it with 60 k. W (V Install our available spare cavity in the BO ring RF = 1. 8 MV) Ø Build a new 60 k. W - 352 MHz SSA identical to a standard tower of the SR amplifiers, using the 160 RF modules of 400 W (BLF 574 transistor) and their dc-dc converters, got back from the upgrade of these amplifiers Ø LLRF & Control a replica of the actual one Ø Increase VRF of the existing plant from 1 MV up to 1. 2 MV P of the existing plant from 1 MV up to 1. 2 MV RF ~ 30 k. W (Pbeam ~ 0) Ø There is free space for the 2 nd cavity in one straight section of the ring and the 60 k. W SSA + LLRF & control inside the booster RF room Infrastructure work Ø Additional benefits : power savings & redundancy in all the other modes of operation Additional benefits Project approved Commissioning beginning of 2018 3

SOLEIL SR RF system Ø En = 2. 75 Ge. V, E = 1. 2 Me. V, Ib = 500 m. A PRF = 600 k. W & VRF : 3 - 4 MV @ 352 MHz Ø 2 cryomodules (CM), each containing a pair of single-cell s. c. cavities (Nb/Cu) Ø Each of the 4 cavities is powered with a 180 k. W solid state amplifier (SSA) Ø Both CM’s are supplied with LHe (4. 2 K) from a single cryogenic plant 4

SR RF system operational results Beam downtime caused by failures from the SR RF over ~ 58 000 running hours in ~ 10 years Equipment Downtime Comments ~ 5 h in 5 short events due to preamplifiers & power combiners ; MTBF ~ 2 years (cumulated by 4 ampli) a) 4 x RF amplifiers ~ 1 10 -4 b) 4 x 500 k. VA PS (230 Vac / 270 Vdc) ~ 4 10 -4 a) + b) = 4 x RF transmitters * ~ 5 10 -4 * MTBF ~ 1 year (cumulated by 4 transmitters) c) 2 x CM’s (4 cavities) ~ 5 10 -4 Tuners, couplers, monitoring pick-ups, vacuum gauges d) 4 x LLRF & control systems ~ 5 10 -4 Wrong interlocks from noise ( filtering & grounding) ~ 23 h in 6 faults from PS control MTBF ~ 2 000 h a) + b) + c) + d) = 4 x RF systems ~ 1. 5 10 -3 e) 1 x Cryo-plant ~ 1. 5 10 -3 ~ 85 h in 6 events (60 h a single one) 2 nd compressor ? ? Last week, He pollution stopped and not yet restarted * Expected << 1 10 -4 with the new generation of transistors (ESRF experience J. Jacob’s talk) and mostly with the new power supplies : ac-dc converters in 2 k. W units redundancy Cryogenics recovery time reduced from 6 down to 3 hours by improving the process control it does not affect anymore machine restart time, after utility failures ; gain of another fact. 2 expected from further upgrade plans : - Installation of a 3 rd 50 m 3 GHe tank Recovery time ~ 1. 5 h, autonomy of ~ 5 hours with - Modifications on the water cooling circuit automatic restart & full compressor redundancy - Revamping of the control system and process 5

Pb with the cavity frequency tuners ü After ~ 2 years of operation, repetitive jams of the cavity freq. tuners, fortunately with relatively small impact on the user runs ü Each cavity has its own tuner which changes its length: double lever & screw-nut assembly, driven by a step motor & a gear box ü Fully housed inside the CM, where it works under vacuum and at cryogenic temperature hard environment for the mechanics ü Try different cures : change of the screw-nut material, threads and backlash without success still jams and degradations !! double lever screw-nut gear box step motor UPGRADED TUNER VERSION 1) Standard screw-nut assembly replaced screw-nut by « planetary roller » screw gear box 2) « Harmonic drive » gear box replaced by « planetary » gear box step motor The four tuners have worked without any trouble for ~ 7 years. Recently, we detected a change in behavior on one of them ; when dismounting, we found that the cage of its screw was broken. The 3 other ones are still working well and no visible wear Dismount one of them for a check. visible 6

Input power coupler (IPC) upgrade Ø Original SOLEIL IPC is a LEP 2 type antenna 200 k. W CW @ 352 MHz Ø Problems of ceramic aging with LEP type IPC’s at ESRF Ø 300 k. W / cav SOLEIL can store 500 m. A using a single CM redundancy cav ≈ In 2011, collaboration agreement with CERN and ESRF develop a new 352 MHz IPC version, based on the LHC design (400 MHz), capable of handling up to 300 k. W. ≈ Six IPC’s were built at CERN and then RF conditioned in the ESRF test-stand up to 300 k. W in transmission and 200 k. W in full reflection, using a copper cavity from CERN The IPC’s were mounted on the CM’s, in situ, situ without removing out of the ring, using a hood with laminar air flow, enclosed within a plastic tent and with slight N 2 gas overpressure inside the cavity “ Clean room ” built on top of the CM ! Remove the « old one » Put on the « new one » 7

IPC multipacting (MP) cure Once mounted on the CM, it took only few days to recondition the IPC’s up to 1. 5 MV with 150 k. W CW full reflection and then we could quickly store up to 500 m. A without any trouble. After ~ 1 week of operation, MP activity at P ~ 110 k. W sometimes vacuum interlocks Re-conditioning during the next shutdowns same scenario : no pb at the beginning of the following run, then a kind of “de-conditioning” after a couple of days. No impact on user runs Pcoup < 110 k. W, compensating with the other cavities PEEK Implementation of a “ bias capacitor ” generating a dc-field at the ceramic window location, aimed at destroying multipacting resonant conditions The multipacting indeed fully disappeared when applying ~ 1 k. V dc and it remained ok in operation bias capacitor burnt PEEK & kapton �� �� ceramic 1) waveguide inner wall 2) inner plate outer plate kapton Bias capacitors damaged by overheating 1) During tests at ESRF Arcing (air side) 2) During operation at SOLEIL And it was still working !! �� no op. impact New EM simulations pointed out dimensioning errors Beg. 2015, replaced + proper temp. controls and since then the 3 operating IPC’s ok (last one to be installed beg 2017). since then the 3 operating IPC’s ok The dirty ceramics and the Cu lips with arc impacts were restored by sand blasting Thanks to the Eric Montessinos team (CERN) 8

Toward storing 500 m. A using a single CM Upgrade of the cavity IPC’s (300 k. W / cav) Modify the waveguide network to combine the power from 2 amplifiers into one cavity Development of a « Magic Switch » P 1+P 2 OR P 1 P 2 Depending on the post configuration CM 2 CM 1 P 1+P 2 P 3+P 4 P 1 P 2 P 3 P 4 Connecting 2 Magic Switches Wave guide network layout to power one or the other CM (300 k. W / cav) from the 4 SSA’s, combined by pairs 9

Upgrade of the 180 k. W SR SSA’s - Excellent operational avaibility and MTBF, but still significant nb of module failures (2 -3 % a year) Matter of maintenance (~ 5 k€ mat. + 3 men. week / year) ; not so bad but still perfectible - « Low cost » upgrade : replace LR 301 transistors (28 V) by BLF 574 XR (50 V) + « module retuning » « Low cost » upgrade Electrical power savings (efficiency : 50 % 60%) compensate the investment cost in < 3 years + More robust transistor & lower thermal stress much less module failures less maintenance + Higher power capability (max Pmod : 310 W 450 W) 500 m. A with only 3 running SSA’s + 7 d. B transistor gain 160 pre-amp modules & their dc PS are got back for the new BO SSA - 3 towers have already been upgraded 4 th one in Oct. 2016 go on at a rate of 1 - 2 towers a year - Not a single failure of a « new » transistor until now (~ 2 years of operation) - Modify the 2. 5 k. W combiners (welded screwed connections) to increase their power capability - Cure the lack of redundancy in the pre-amplification stage develop a “combiner-divider” combiner-divider 24 W 3 W 1 1 Pre-ampli 300 W 3 W 300 W 2. 4 k. W 80 300 W 80 24 k. W 190 k. W Present config : each pre-ampli drives 80 modules; if one of them fails the amplifier is stopped Upgrade 190 k. W Thanks to the combiner-divider, combiner-divider the failure of a pre-ampli does not affect the functioning anymore 10

Other activities on SSA’s § Collab. with LNLS (Brazil) 2 x 50 k. W SSA’s @ 476 MHz (400 W BLF 574) § ESRF (transfer techno Elta) 7 x 150 k. W SSA’s @ 352 MHz (700 W BLF 578) § 500 MHz SSA’s for Thom. X (50 k. W) and SESAME (80 k. W) 50 k. W SSA for Thom. X - Fully modular PS : 230 Vac / 50 Vdc (6 dissipators x 16 mod) in 2 k. W units of 96% efficiency and voltage remote control re-optimize efficiency * for any operating power : 56% (overall) @ Pmax & 50% @ 0. 6 Pmax Tower config. & dc/dc converters + rectifier * + 10 pts in efficiency lead to electricity savings over 10 years ≈ full ampli cost - Change to cabinet configuration keeping the exchangeability at the lowest level (elementary module) optimum modularity/redundancy and spare inventory management - Improved control and supervision Ethernet SNMP PLC DPA 80 k. W SSA for SESAME (10 dissipators x 16 mod) MUX : analogic comparators MUX & multiplexers + a µcontroller which monitors all data from a dissipater (16 mod & their PS) + CPLD for the interlocks 11

Other activities on SSA’s q We’ve completed the Thom. X and first SESAME SSA’s; the 3 other ones for SESAME are being built Thom. X and first SESAME SSA’s; by Sigma. Phi Electronics (SPE), the SOLEIL licensee. Sigma. Phi Electronics Wa. CCo q The 150 k. W-500 MHz SSA is already in the SPE catalogue The 150 k. W-500 MHz SSA 2 x 75 k. W (2 x 8 dissipaters of 16 modules) combined by means of a wave guide to coaxial combiner, the Wa. CCo q 1. 3 GHz SSA for LUCRECE (R&D for LUNEX 5) SPE has already built SSA’s at 1. 3 GHz (Pmod ~ 200 W) 9 x 10 k. W for ELBE* & 1 x 16 k. W for b. ERLin. Pro SOLEIL - SPE 20 k. W - 1. 3 GHz SSA , Pmod > 400 W (LDMOS vs Ga. N transistor) * Hartmut Büttig’s poster 60 long bunch 40 short bunch 20 V [MV] q R&D on VHF e-gun for LUNEX 5 ~ 100 k. W SSA (@ 176 MHz ? ) q SOLEIL upgrade towards DLSR and VSR DLSR : 6 & 7 BA / cell lattice e ~ 200 pm VSR : alternately long (10 ps) & short (0. 1 ps) bunches along the train, obtained by adding 2 harmonic systems, h 1 = 5 and h 2 = 5. 5 Replace one of the actual 352 MHz CMs by another one containing a pair of sc cavities of each frequency, either passive or powered with 10 k. W SSA’s @ 1. 76 and 1. 94 GHz ? 0 0 1 2 3 4 5 6 -20 short -40 -60 352 MHz (1) 1. 76 GHz (2) 1. 936 GHz (3) SOMME 1+2+3 t [ns] 12

Summary - Conclusions q The 352 MHz RF systems of SOLEIL are quite innovative with the use of HOM free sc cavities and SSA’s, both developed in house. After ~ 10 years of operation they have demonstrated excellent reliability and flexibility. q The difficulties encountered with the sc cavity frequency tuners were quickly solved by improving the original tuning device. q Cavity IPC’s of higher power capability (300 k. W) have been developed in collaboration with CERN and ESRF 3 out of the 4 cavities are equipped with the new IPC’s + bias voltage for coping with recalcitrant multipactor levels; the last one will be implemented beg. of 2017. Modifications of the waveguide network will give the possibility of combining 2 amplifiers and thus power each cavity of a CM with 300 k. W Storage of 500 m. A using a single CM. q A special emphasis is put on the success of the SSA’s, developed by SOLEIL at 352 MHz; they have demonstrated that they can advantageously replace the vacuum tubes in such an application (extreme modularity, absence of HV, very low phase noise, …). very low phase noise q R&D carried out at SOLEIL has allowed to improve the original 352 MHz design (overall efficiency > 60 %) and extend it to other frequencies. It has now reached maturity, being adopted by many %) other facilities and taken up by the industry for applications ranging from 80 MHz up to 1. 5 GHz. q The French company SIGMAPHI ELECTRONICS (SPE) is SOLEIL licensee since Dec. 2013. § For SESAME : 4 x 80 k. W - 500 MHz SSA’s, 1 st one built by SOLEIL, 3 other ones by SPE. § SOLEIL & SPE 20 k. W - 1. 3 GHz SSA (LDMOS vs Ga. N) for LUCRECE/LUNEX 5. q Future projects : § VHF (176 MHz ? ) e-gun for LUNEX 5 § SOLEIL upgrade towards « DLVSR » SC harmonic RF systems, 1. 75 GHz & 1. 94 GHz 13

Acknowledgements SOLEIL RF and LINAC group Patrick MARCHAND Jean-Pierre POLLINA Robert LOPES Fernand RIBEIRO Massamba DIOP Cyril MONNOT Julien SALVIA Helder A. DIAS Renaud CUOQ Marc LOUVET Rajesh SREEDHARAN Jocelyn LABELLE Jean-Pierre BAETE Sylvain PETIT Many thanks also to all the members of CERN and ESRF who were involved in the fabrication and/or tests of the new IPC’s.